Visit our website:
http://www.france24.com
Like us on Facebook:
https://www.facebook.com/FRANCE24.English
Follow us on Twitter:
https://twitter.com/France24_en
Category
đ
NewsTranscript
00:00We're going to turn now to the co-founder of the Indian Software Products Industry Roundtable,
00:06the iSpirit Foundation, Sharad Sharma. Thanks for being with us. You've known the age of the floppy
00:14disk, and now you know the age of artificial intelligence. And what's been interesting in
00:19the last minutes listening to Kemi is we seem to be having this sort of schizophrenic kind
00:24of conversation. On the one hand, we're like, this is great. How fast can we go to make this stuff?
00:32And on the other hand, we're wondering, does humanity know what it's doing? Are we
00:37creating the tools of our own downfall? Does AI keep you up at night?
00:42It does. And the reason both of those aspects are important is because if you play AI right,
00:50it will lift many Indians out of poverty. It will give them better health. It will let
00:57our students study better and learn better. But on the other hand, if we don't do AI well,
01:04and we don't have all the safety provisions that we need, then we will have children who
01:10will be affected by AI pornography, AI gambling, AI gaming, and we'll have a lost generation.
01:16And the Indian lost generation is a big generation. We had 23 million babies born last
01:20year. We have the largest number of under 15-year-olds in the world. So we have to worry
01:28about both the opportunity and the threats at a scale that no other country has to worry about.
01:34And the services sector powers the Indian economy right now.
01:39I would say yes.
01:40And so is this going to be a job destructor?
01:43It could be. And on the other hand, it could enable people to fulfil their best potential,
01:49right? And so you could have many outcomes in the future. And I think the only way we can play this
01:58is that luckily for us, India did have a revolution in digital world, as some of you may
02:04know. India has been very successful in the last 15 years in creating digital infrastructure that's
02:11brought everybody into the system. So India is a unique country because it's data rich even before
02:17it's economically rich. And as you know, AI is powered by data, right? So our health data can
02:24inform how people can be healthier in the future. Our education data can inform how people can
02:30study and learn more effectively. Our financial data, because we've had a huge revolution in
02:36financial services in India, that can inform how small businesses can be given loans in a way that
02:42is not possible to give right now. So AI offers us, the next generation of AI will offer us many
02:49opportunities. Why? Because that would be built on real data. Currently, AI is built on public data.
02:56That's also real, but it's available to everybody. Now, tomorrow's AI...
03:00Some say available thanks to stealing people's data.
03:03Perhaps, and therefore the next set of data is going to be about health records of hundreds of
03:08millions of people. Can you use it for AI in a privacy preserving way? Can you use people's
03:15spending patterns, that data in a privacy preserving way? Can you do that for education
03:21in a privacy preserving way? And India has built the deep, we call it digital public infrastructure
03:27that allows this to happen in a privacy preserving way. And because of that,
03:32India is going to unlock continental data sets in a way that no other country is able to do,
03:38or no other place is able to do right now. So the future of AI is actually going to depend
03:43on those kinds of data sets and the kind of AI that emerges from it.
03:48And that's going to determine what world will see as AI in the future.
03:53How do you prevent it from becoming the dream of any police state,
03:58like a dystopian movie like Minority Report?
04:01In fact, that is a very important issue. So just the way we talk about security,
04:05the most important word in the future is going to be privacy. And most people don't appreciate
04:10that the first country to make privacy a fundamental right was in India. Our Supreme
04:18Court gave a 9-0 judgment, which cannot be overturned easily in the future, saying privacy
04:24is a fundamental right. We have the biggest system for private data sharing. So in Europe,
04:30you have this idea that personal data, private data can only be shared with consent. I don't
04:35know, many people may not know that India's consent system is an electronic system. It's open,
04:42it's revocable, it's granular, it's auditable, it's notified and it's secure, organs we call it.
04:48It runs at 120 million consents a year today. It will end this year with 600 million consents
04:57this year. So therefore, the scale at which we are able to build privacy into the techno legal
05:04system, what is India's specialization is that we use technology and policy, public technology,
05:12public policy, coming together to protect Indian citizens from overreach by companies
05:18or overreach by the state. Yeah, because we saw with DeepSeek, this large language model put out
05:25on the cheap by the Chinese, as we heard Peter O'Brien say, there are reasons to hope, but also
05:31reasons to worry when you type in things like Tiananmen Square, nothing comes out. How do you
05:36make it so that if you type in IOTA, for instance, which is something contested in India over what
05:40happened there, that your large language model isn't going to have a bias? Look, even in DeepSeek,
05:46if you install it on your own computer, you can get whatever information that you want,
05:53because all the controls are in the layers above, which can be stripped away.
05:57Remember, all these AI models are eminently shareable. One of the misconceptions that we
06:03had with AI is that we could put a regime that would prevent proliferation of AI models.
06:10That's not true, because AI model after it's been trained in the end is set of model weights,
06:16and those model weights can be put on a USB and shared with everybody. That's what you can do
06:21with DeepSeek as well. So when you install DeepSeek, don't use it as a service, but install
06:27it on your own computer and run it, then it will give you the stuff that it has been truly trained
06:34on without the censorship that has been added on later.
06:37So, Sharad Sharma, we heard our correspondent there allude to it. You have the Indian Prime
06:44Minister Narendra Modi coming to Paris. There's this nail biting over what will be in the final
06:49joint declaration when it comes to AI safety. Then Narendra Modi goes on to Washington.
06:55And we know that these days, Donald Trump likes to rip up international treaties.
07:00At the end of the day, is it the global regulation that wins,
07:04or is it the move fast and break things crowd that win the argument?
07:09I think there are two competing views on AI. One is that, will it be a source of state power?
07:19The Chinese and the Americans would like it to be a part of state power. And if indeed that
07:26is the winning paradigm, then India, and I would say even France, would want to have strategic
07:32autonomy in AI, because we would not like to be client states of either side. The second view is
07:40that no, AI will proliferate. It will be hundreds and thousands of fireflies lighting up the sky.
07:48That's what they promised with the internet. It didn't happen that way. Instead,
07:52concentration of power we have in Silicon Valley.
07:54Therefore, if you hear the India stack story, or the DPI story, that is the story of India.
08:01India has taken, what was internet? Internet was flow of data packets, flow of messages,
08:08email messages, TCP, IP, if you know protocols, SMTP, and the last one is the flow of web content.
08:17India made flow of people, money, and information as part of public infrastructure,
08:24digital public infrastructure. Companies like Microsoft and Alphabet,
08:29the parent company of Google, run our world. No, in India, our payment system, which is a
08:35digital public infrastructure, Google is a big player, but it runs it on Indian terms.
08:41See, this is what I think is the big learning for Europe, that if you really want to shape market
08:48behavior, you cannot in the digital world shape it with one hand. Think of it as you are shaping
08:55clay on a potter's wheel. Can you do it with one hand? You can't. You need two hands. The two hands
09:00are public policy and public technology. Now, Europe used to be the epicenter of public
09:06technology. That has dissipated. US used to have lots of public technology. I would submit to you
09:12that the largest pool of public technologies today are sitting in Bangalore, the city I am from.
09:18And so, therefore, today, if you really want to get good outcomes, you need two hands,
09:24public technology and public policy. That is what will shape the digital and the AI future of the
09:31world. And that is called techno-legal regulation. That is the contribution of India to the discourse
09:37about what the future of digital and future of AI will be in the coming year. And if J.D. Vance,
09:43who's representing the United States, doesn't want to sign on, you're saying the Indians and the
09:48French and whoever else will do it. India's digital revolution has managed to prevent Alibaba from
09:56running our systems and has prevented the big tech in the US from running our systems. And we
10:01have done it in a way, for example, where some of our systems like MOSIP are in 26 countries. 943
10:09million citizens are signed up to use it. Why? Because it's not weaponizable. Privacy is built in.
10:17It has techno-legal regulation. So, there is a new philosophy that is emerging. Why am I here in
10:24France? Because originally this philosophy was the French philosophy. It used to be called digital
10:30commons and data commons. However, that philosophy was overturned by the way from Silicon Valley.
10:42In India, we learned from that. We figured out how to keep it alive and run it at scale and we
10:48are doing it. And therefore, that has become the last bastion where you can see this working,
10:54where the values of the internet are still operative and will probably become stronger
10:59as we go forward. Stronger as we go forward. Sharad Sharma, I want to thank you so much for
11:03being with us here on France 24. Thanks again, Kemi Knight with that full coverage, which again
11:08continues through to next Tuesday with that AI summit coming up. Stay with us, top of the hour,
11:12more news and then the France 24 debate. And how does Ukraine these days handle Donald Trump?
11:19That's the question we've been posing to you on the hashtag F24 debate.